On Fri, Dec 5, 2008 at 9:22 PM, Robert Bullock <bobb@prostyle.com> wrote:
> Hi Solarwind,
>
> I think this is a great project and I am really impressed that someone 17 is
> taking it on out of general interest. You remind me of me, although
> unfortunately when I was your age, PIC's were well in the future.
>
> I suspect you are a little discouraged by all the nonsense that has been
> posted in reply to your initial question that has nothing to do with the
> issue at hand.
>
> First step is to determine what families of MCU you have access to a good C
> compiler for, at a cost you can afford (free most likely). A compiler from
> any of the major players would be fine for this. Hitech, Microchip or IAR
> for example. If any offer a trial version, it can't be crippled by code size
> as you will need the full abilities of the compiler. Time limited, or maybe
> an older version, would be ok.
>
> The most important aspect of the C compiler in your case will be that it
> supports 32 bit floating point data type.
>
> There are several factors to consider when selecting the best MCU.
>
> 1) As already stated, access to a good C compiler.
> 2) Large enough code space and ram.
> 3) Power consumption.
> 4) Cost.
> 5) Required peripherals.
> 6) Availability of useful support libraries.
>
> Fortunately in your case, you have it much easier than designers of
> commercial products as issues like cost and power consumption are far less
> important when you are making one of a product, and its primary purpose is
> for fun.
>
> In regards to power consumption, many of Microchips MCU's offer what they
> call nanoWatt features that are aimed at helping design systems for very low
> power usage. In your case, I would think the ability to change the system
> clock frequency on the fly would be the most important feature to look for.
> For the most part, power consumption is directly related to the frequency at
> which the part is switching at. So when the calculator is not processing a
> calculation, and is basically waiting on user input, and running the
> display, it can be running at a very low frequency, such as 72 kHz. As soon
> as you have a computation to perform, you switch the processor to the
> highest frequency, do the calculation, and switch back down.
> You may even be able to put the processor to sleep, using a wakeup on button
> press approach.
>
> I would look at selecting a processor with hardware multiply and divide as
> this may be required to complete some of the high order scientific functions
> quickly enough. It also may have other potential savings in code size and
> power consumption.
>
> The obvious problem you face is that you do not know how much you will need
> in many of these areas. What I always like to do is to start with an MCU
> that I know is fast enough, and has enough ram and flash, for my project.
> Experience helps with this process. In your case, I would simply start with
> the largest, fastest MCU that meets the other requirements outlined above.
>
> If it was me doing this, I would look very hard at the PIC24H series of
> parts. I would also use the Microchip compiler. You can use any editor you
> wish to write the software, and for your situation, differences in the
> development IDE would not be too important.
>
> Bob Bullock
>
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

Thanks a lot! The compiler here is not really an issue for me. It is
trivial to snap crack and pop the demo pro version of it.

-- 
..::[ solarwind ]::..
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist